AI Skill Center
The Skill Center manages structured knowledge cards in SKILL.md format. Skill cards are knowledge packages that AI Agents can reference during penetration testing (e.g., web evasion handbooks, cloud/container assessment playbooks, mobile reverse-engineering methods), orchestrated by the platform and exposed to the Agent side.
Navigate to "AI" → "Skill Library" in the left sidebar (requires skill:view or mcp:view permission).
Skill Data Model
| Field | Description |
|---|---|
| skillKey | Unique identifier (e.g., advanced-web-evasion) |
| Name / Category | Display name and category |
| Description | Usage summary that helps the Agent decide whether to load it |
| Content | SKILL.md Markdown body (with DOMPurify-sanitized preview) |
| Enabled | Disabled skills are not served to Agents |
| Sort Order / Version | Display order and version; built-in skills are flagged isBuiltin |
Category System
Skills are organized into 11 categories, filterable by category plus debounced keyword search:
BASE / WEB / SAST (code auditing) / API / CLOUD / AI (AI/LLM apps) / MOBILE / COLLAB / VULN / ORCHESTRATION / CUSTOM
Common Operations
Create & Edit
Click "New Skill" to open the drawer editor. Fill in skillKey, name, category, and the Markdown body with live preview.
Enable / Disable
Toggle the switch on each card to control whether the skill takes effect for AI Agents.
Export & Distribution
- One-click Export: Bundle all enabled skills as
skills.zip, ready to drop into an Agent's skill directory (e.g., Claude Code's.agents/skills/) - Single Download: Download a specific skill's
SKILL.mdfile - Copy: Copy the skill's Markdown content to the clipboard
Views & Browsing
Grid/list dual views with pagination.
Related Documentation
- MCP Overview — how skills are exposed to AI Agents via MCP